Abstract
In a seminal article, Yin (1969) reported that the recognition of face pictures was disproportionately affected by a 180 deg rotation in the image plane from the normal, upright viewing condition. This phenomenon is now commonly called the Face Inversion Effect (FIE). There is now agreement amongst most face recognition researchers that the FIE arises from a greater difficulty to perceptually encode inverted face information (e.g. Farah et al., 1998). Recent studies have therefore examined more closely these encoding differences. The exact nature of these differences remain so far largely unknown (e.g., Rossion & Gauthier, in press).
In order to throw some light on the FIE, we used Bubbles, a technique developed to reveal diagnostic visual information (Gosselin & Schyns, 2001), in the context of a face priming paradigm. Our stimulus set (prime and target) comprised 32 faces (8 individuals * 2 genders * 2 expressions). A trial consisted in the presentation of a «bubbled» face (i.e., a face sparsely sampled in spatial location space) for a duration of 100 ms (the prime) followed by the presentation of a white noise mask for a duration of 15 ms, and trailed by the re-presentation of the prime stimulus but unaltered this time (the target). On half the trials the prime was inverted; the target always appeared upright. Eight subjects viewed 2,560 trials (divided in 4 blocks). We computed separate multiple regressions of the bubble masks and response times for the upright and inverted conditions, and performed statistical analyses on the regression coefficients. When the prime was upright, the left eye, the mouth, and a region revealing a small part of the left border of the face produced a significantly larger priming effect than other regions; no region reached significance when the prime was inverted. We discuss the implications of these results for the configural-but-not-featural-disruption explanation of the FIE.